The success of spermatid microinjection has generated many concerns. In particular, there is a lack of appropriate methodology for the isolation of large homogeneous populations of spermatids, with minimum loss of viability, from the testicular tissue of azoospermic men. Here we have compared two different isolation methods -- velocity sedimentation under unit gravity (VSUG) combined with discontinuous Percoll centrifugation (DPC), and separation with fluorescent-activated cell sorter (FACS) using light in the visible range -- to determine the most suitable method for the isolation of spermatids. Total mixed cell count/gram of testicular parenchyma was significantly higher in obstructive azoospermic men compared with non-obstructive azoospermic men (P < 0.001). The results of the comparison showed that in obstructive azoospermic patients the difference in the yields of primary spermatocytes produced by the two techniques was not significant, but for round and elongating spermatids the FACS separation proved to be the better method (P < 0.001). Similarly, in non-obstructive azoospermic patients, FACS separation proved to be superior, giving increased yields of primary spermatocytes and round and elongating spermatids compared with VSUG combined with DPC method (P < 0.001). More than 99 % of the separated cells retained their viability after FACS separation. As large homogeneous populations of viable spermatids can be separated with FACS in a relatively short period of time, FACS separation is the most suitable method for the isolation of spermatids from testicular biopsy tissue.